Unfortunately there is no alternative to this yet with MySQL Cluster. I
believe there is work in progress to improve blob handling but I cannot
say whether this would help in your application. I would recommend
alternative methods such as storing the blobs as files on SAN/NFS and
the filenames for the blobs in the database. If you really need the
data to come via. the database you could use a UDF to pull the blob data
from the file based on the filename in the record and return it as part
of a row.
On a related note if you are using mysqld for your API nodes it would be
a good idea to wait for 7.0.8 to be released. There are a few leaks and
late memory release problems related to blobs fixed in mysqld in this
version.
"I would recommend alternative methods such as storing the blobs as files on
SAN/NFS and the filenames for the blobs in the database."
Can you help me with the procedure of how to implement it?
" If you really need the data to come via. the database you could use a UDF
to pull the blob data from the file based on the filename in the record and
return it as part of a row."
Can you please elaborate its procedure a bit?
"On a related note if you are using mysqld for your API nodes it would be a
good idea to wait for 7.0.8 to be released. There are a few leaks and late
memory release problems related to blobs fixed in mysqld in this version."
Since 7.1.X series is currently in use, is there any improvement in the BLOB
storage mechanism. The architecture we are currently implementing is still
consuming a lot of RAM (even after using BLOB storage).
---
Regards
Puneet Sachdev
Software Developer
Northern Lights Studio Pvt. Ltd.
E-mail: puneet.sachdev@stripped
